NFS: Always wait for I/O completion before unlock

NFS attempts to wait for read and write completion before unlocking in
order to ensure that the data returned was protected by the lock.  When
this waiting is interrupted by a signal, the unlock may be skipped, and
messages similar to the following are seen in the kernel ring buffer:

[20.167876] Leaked locks on dev=0x0:0x2b ino=0x8dd4c3:
[20.168286] POSIX: fl_owner=ffff880078b06940 fl_flags=0x1 fl_type=0x0 fl_pid=20183
[20.168727] POSIX: fl_owner=ffff880078b06680 fl_flags=0x1 fl_type=0x0 fl_pid=20185

For NFSv3, the missing unlock will cause the server to refuse conflicting
locks indefinitely.  For NFSv4, the leftover lock will be removed by the
server after the lease timeout.

This patch fixes this issue by skipping the usual wait in
nfs_iocounter_wait if the FL_CLOSE flag is set when signaled.  Instead, the
wait happens in the unlock RPC task on the NFS UOC rpc_waitqueue.

For NFSv3, use lockd's new nlmclnt_operations along with
nfs_async_iocounter_wait to defer NLM's unlock task until the lock
context's iocounter reaches zero.

For NFSv4, call nfs_async_iocounter_wait() directly from unlock's
current rpc_call_prepare.

Signed-off-by: Benjamin Coddington <bcodding@redhat.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>

1 files changed, 9 insertions, 0 deletions

diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index c52f72c86940..dbfc7574fab0 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -5784,6 +5784,7 @@ struct nfs4_unlockdata {
         struct nfs_locku_res res;
         struct nfs4_lock_state *lsp;
         struct nfs_open_context *ctx;
+        struct nfs_lock_context *l_ctx;
         struct file_lock fl;
         struct nfs_server *server;
         unsigned long timestamp;
@@ -5808,6 +5809,7 @@ static struct nfs4_unlockdata *nfs4_alloc_unlockdata(struct file_lock *fl,
         atomic_inc(&lsp->ls_count);
         /* Ensure we don't close file until we're done freeing locks! */
         p->ctx = get_nfs_open_context(ctx);
+        p->l_ctx = nfs_get_lock_context(ctx);
         memcpy(&p->fl, fl, sizeof(p->fl));
         p->server = NFS_SERVER(inode);
         return p;
@@ -5818,6 +5820,7 @@ static void nfs4_locku_release_calldata(void *data)
         struct nfs4_unlockdata *calldata = data;
         nfs_free_seqid(calldata->arg.seqid);
         nfs4_put_lock_state(calldata->lsp);
+        nfs_put_lock_context(calldata->l_ctx);
         put_nfs_open_context(calldata->ctx);
         kfree(calldata);
 }
@@ -5859,6 +5862,10 @@ static void nfs4_locku_prepare(struct rpc_task *task, void *data)
 {
         struct nfs4_unlockdata *calldata = data;
 
+        if (test_bit(NFS_CONTEXT_UNLOCK, &calldata->l_ctx->open_context->flags) &&
+                nfs_async_iocounter_wait(task, calldata->l_ctx))
+                return;
+
         if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
                 goto out_wait;
         nfs4_stateid_copy(&calldata->arg.stateid, &calldata->lsp->ls_stateid);
@@ -5910,6 +5917,8 @@ static struct rpc_task *nfs4_do_unlck(struct file_lock *fl,
          * canceled lock is passed in, and it won't be an unlock.
          */
         fl->fl_type = F_UNLCK;
+        if (fl->fl_flags & FL_CLOSE)
+                set_bit(NFS_CONTEXT_UNLOCK, &ctx->flags);
 
         data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
         if (data == NULL) {